Multilayered Reprogramming in Response to Persistent DNA Damage in C. elegans

Summary

DNA damage causally contributes to aging and age-related diseases. Mutations in nucleotide excision repair (NER) genes cause highly complex congenital syndromes characterized by growth retardation, cancer susceptibility, and accelerated aging in humans. Orthologous mutations in Caenorhabditis elegans lead to growth delay, genome instability, and accelerated functional decline, thus allowing investigation of the consequences of persistent DNA damage during development and aging in a simple metazoan model. Here, we conducted proteome, lipidome, and phosphoproteome analysis of NER-deficient animals in response to UV treatment to gain comprehensive insights into the full range of physiological adaptations to unrepaired DNA damage. We derive metabolic changes indicative of a tissue maintenance program and implicate an autophagy-mediated proteostatic response. We assign central roles for the insulin-, EGF-, and AMPK-like signaling pathways in orchestrating the adaptive response to DNA damage. Our results provide insights into the DNA damage responses in the organismal context.

Graphical Abstract

Highlights

• •Proteome responses to persistent DNA damage correlate with starvation and aging
• •Proteostatic shift reduces ubiquitin proteasome and chaperones and relies on autophagy
• •Metabolic adaptations to DNA damage reduce fatty acid synthesis
• •Insulin-, EGF-, and AMPK-like signaling pathways respond to UV-induced DNA damage

Edifizi et al. provide a comprehensive proteomics, phosphoproteomics, and lipidomics analysis of the response to persistent DNA damage in a metazoan organism. Proteostasis shifts toward autophagy, fatty acid metabolism is attenuated, and the insulin-, EGF-, and AMPK-like signaling pathways form the center of the response network.

Introduction

DNA damage accumulation is a driving factor for the aging process. Congenital syndromes that are caused by mutations in genome maintenance pathways are characterized by accelerated aging and premature onset of aging-associated diseases (Vijg and Suh, 2013). The role of unrepaired DNA lesions in cancer development and (premature) aging is particularly well exemplified in nucleotide excision repair (NER) deficiency syndromes (Edifizi and Schumacher, 2015). NER removes bulky lesions such as UV-induced cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts (6-4 PPs). Two distinct NER sub-pathways recognize the lesions: global-genome (GG) NER scans throughout the entire genome, and transcription-coupled (TC) NER initiates repair when RNA polymerase II stalls at a lesion. Although mutations in GG-NER lead to highly elevated skin cancer susceptibility in xeroderma pigmentosum (XP) patients, TC-NER-deficient Cockayne syndrome (CS) patients suffer from growth and mental retardation and premature aging (Ribezzo et al., 2016).

Given the highly complex NER phenotypes in human patients and respective mouse models, we have employed the nematode C. elegans as a metazoan model to better understand the consequences of unrepaired DNA damage. Mutations in the GG-NER gene xpc-1 lead to genome instability in proliferating cells, which in adult nematodes are restricted to the germline, whereas TC-NER-deficient csb-1 mutants cease developmental growth when exposed to UV irradiation (Lans et al., 2010, Mueller et al., 2014). Thus, GG-NER defects are linked to genome instability in proliferating cells, a hallmark of cancer development in humans, whereas TC-NER defects mirror the growth defects and accelerated decline in tissue functionality associated with CS (Edifizi and Schumacher, 2015). We have previously employed the nematode NER mutants to gain insights into the response mechanisms to persistent DNA damage during development and aging. We established that the insulin/insulin-like signaling (IIS) effector DAF-16 counteracts DNA damage-driven aging by elevating tolerance to persistent DNA damage (Mueller et al., 2014). Strikingly, NER-deficient csb-1;xpa-1 double mutant and ercc-1 mutant mice that display growth defects and accelerated aging show dampening of the IIS-equivalent somatotropic axis (Niedernhofer et al., 2006, van der Pluijm et al., 2007). The main signaling components of the somatotropic axis, the growth hormone receptor (GHR) and insulin-like growth factor-1 receptor (IGF-1R), are downregulated in response to persistent transcription-blocking lesions (Garinis et al., 2009), suggesting highly conserved DNA damage response mechanisms during nematode and mammalian development and aging.

Given the role of unrepaired DNA lesions in progeroid syndromes and the contribution of accumulating DNA damage to the aging process, we devised a study aimed to gain a more comprehensive understanding of the response mechanisms to persistent DNA lesions on the organismal level. We used xpc-1;csb-1 mutant worms that are defective in both GG-NER and TC-NER, thus leading to complete inability to remove UV-induced DNA lesions resulting in persistent DNA damage after UV treatment (Mueller et al., 2014). We used the UVB irradiation in order to induce helix-distorting CPDs and 6-4 PPs throughout the tissues of the animal. We analyzed proteome, phosphoproteome, and lipidome alterations in response to UV irradiation of NER-deficient animals. On the proteome level, we found similarities between the response to UV irradiation in NER-deficient animals and proteome alterations during aging (Walther et al., 2015) and in the response to starvation (Larance et al., 2015), both of which are regulated through the IIS pathway (Depuydt et al., 2014). Next, we devised a comprehensive signaling response network to DNA damage by integrating proteome and phosphoproteome changes upon persistent DNA damage. Our analysis thus provides insights into the animals’ physiological adaptations to unrepaired DNA damage. Furthermore, analyzing the lipidome, we identified metabolic alterations that indicate a shift to somatic preservation in response to DNA damage. Consistent with the metabolic adjustments, we observed a reduction in proteins functioning in carbohydrate, amino acid, and lipid metabolism that resemble metabolic changes observed upon starvation and during aging. Mechanistically, we determined an important role of autophagy and AMPK signaling in the maintenance of tissue functioning amid persistent DNA damage.

Results

We applied mass-spectrometry-based quantitative proteomics to completely NER-deficient xpc-1(tm3886);csb-1(ok2335) double-mutant C. elegans. We analyzed synchronized worms at the first larvae stage (L1), 6 hr after UVB or mock treatment. Proteins were digested in solution followed by peptide identification and quantification by liquid chromatography and tandem mass spectrometry (LC-MS/MS) (Figure 1A). In total, more than 7,500 proteins were quantified at a false discovery rate (FDR) of less than 1% at the protein and peptide spectrum match level, of which more than 5,000 proteins were quantified between UV and untreated conditions at least in two out of three biological replicates. Excellent reproducibility (r > 0.95 for biological replicates) was determined with the Pearson correlation coefficient (r). Hierarchical clustering revealed strong segregation of different conditions indicating distinct proteomic changes and high data quality that allows for systematic analysis (see also Figure S1). By using a two-sided t test and correcting for multiple testing by estimating the FDR to 5% with a permutation-based algorithm, we identified about 1,000 significantly differentially expressed proteins, of which more than 550 proteins were more than 2-fold altered between UV and untreated conditions (Table S1).

Figure 1.

Proteome Analysis of the DNA Damage Response in NER Deficient C. elegans

(A) Experimental workflow. xpc-1;csb-1 double-mutant L1 larvae were treated with 100 mJ/cm² UV and proteome-analyzed by LC-MS/MS.

(B) Significantly increased (red; >1.5-fold up) or decreased (blue; >1.5-fold down) proteins (FDR < 5%) in the different subcellular compartments (see Table 1 for details on clusters).

(C) Volcano plot of the proteins detected upon UV treatment, including significantly (FDR < 5%) increased (red) and decreased (blue) proteins.

(D) GO categories of human (orange) and C. elegans (blue) annotated proteins.

To systematically analyze protein changes upon persistent UV lesions, we used Gene Ontology (GO) classification as well as UniProt (January 2016 release) and the C. elegans portal WormBase (version WS246). Abundance changes of significantly regulated proteins were observed in most of the subcellular compartments (Figure 1B). In the volcano plot, the log₂ ratio of UV-treated worms to untreated worms for each protein group is plotted against the respective –log₁₀ p value (Figure 1C). In order to improve the annotations of C. elegans proteins and obtain insights into potential functions, we used BLAST search results (e < 10⁻⁴) of well-annotated human and mouse proteins. We used GO, Kyoto Encyclopedia of Genes and Genomes (KEGG), and gene set enrichment analysis (GSEA) annotations provided by the UniProt database for C. elegans protein entries and the corresponding human orthologs raising protein annotations from ∼35% to ∼62% (Figure S2).

We used 1D enrichment to identify groups of proteins that are involved in identical pathways, carry similar PFAM domains, or localize in the same compartment (e.g., categorical annotations). We visualized significantly regulated groups (Benjamini-Hochberg FDR < 0.02) by plotting the mean log₂ ratio of UV-treated to mock-treated worms for all proteins with the particular categorical annotation against the enrichment score (Figure 1D). Categories grouping proteins related to nuclear mechanisms and synaptic machinery showed a positive enrichment score, whereas categories related to protein synthesis and cellular metabolic processes showed a significant negative score (Table S2). Overall, the systematic analysis indicates widespread changes of protein levels upon UV-induced DNA damage in C. elegans.

Upregulated Protein Clusters upon Genotoxic Stress

Proteins belonging to the categories related to nuclear mechanisms such as chromatin remodelers, regulator of transcription, protein-DNA complex, and structures of the nuclear pore showed clear upregulation, consistent with chromatin remodeling modulating replication and transcription in response to DNA damage. In addition, the increased expression of members of the synaptic machinery and G protein signaling partners, belonging to plasma membrane and extracellular space categories, suggests that signals are released from genotoxically compromised cells that mediate the adaptation to the damage.

The significantly enriched upregulated proteins belonging to the nuclear GO category (Figure 1B; Table 1) includes chromatin remodelers (CHD-7, BAF-1, SWSN-4, SNFC-5, and LMN-1), transcription regulators (HMG-1.2, RTFO-1, STA-1, NONO-1, EMB-5, SPT-4, HCF-1, and SMK-1), and histone post-translational modifiers (SPR-5, HIL-2, HTZ-1, and HDA-3) associated with the epigenetic control of gene expression. The chromatin-associated proteins BAF-1, SWSN-4, and HCF-1 were previously shown to interact with the IIS effector DAF-16 to remodel chromatin and activate transcription (Li et al., 2008, Riedel et al., 2013). Other transcription factors mediate specifically the response to DNA damage and oxidative stress (SMK-1) (Wolff et al., 2006) or play a role in the UV-induced DNA damage response in mammalian cells (NONO-1) (Alfano et al., 2016). The upregulated proteins include transcription elongation, pre-mRNA processing proteins, and ribonucleoprotein (RNP) (Table 1), in line with changes in spliceosome organization and the post-translational modifications of splicing factors, recently implicated in the DNA-damage response (Tresini et al., 2015).

Table 1.

Most Significantly Overrepresented Cluster of Proteins (FDR < 5%) Increased (Upper Section) or Decreased (Lower Section) in Abundance in xpc-1;csb-1 Double Mutants upon UV Irradiation

	Protein Name	Biological Function	Fold Change
Proteins Significantly (FDR < 5%) Increased in Abundance in UV-Treated versus Untreated xpc-1;csb-1 Double Mutants
Nuclear
Histones	spr-5	lysine-specific histone demethylase 1	1.7
	rfp-1	E3 ubiquitin-protein ligase mediating monoubiquitination of histone H2B	1.99
	htz-1	histone H2A	2.08
	hda-3	histone deacetylase	2.3
	hil-2	histone H1.2	2.56
Chromatin organizers	swsn-4	SWI/SNF nucleosome remodeling complex component	1.61
	chd-7	chromodomain and helicase domain protein	1.65
	vrk-1	Ser/Thr kinase regulating the association of baf-1 with chromatin and nuclear membrane proteins	1.66
	lmn-1	lamin-1, major component of the nuclear lamina	1.67
	emr-1	emerin homolog, involved in chromosome segregation and cell division	1.85
	baf-1	barrier-to-autointegration factor, essential role in NE formation	2.26
	lem-2	LEM protein, involved in chromosome segregation and cell division	2.5
	snfc-5	SNF chromatin remodeling complex component	2.53
Chromosome cohesion	smc-3	structural maintenance of chromosomes protein 3	1.94
	coh-1	cohesin complex subunit	2.52
	scc-3	cohesin complex subunit	3.1
Regulators of transcription from RNA polymerase II promoter	math-33	ubiquitin carboxyl-terminal hydrolase	1.62
	smk-1	suppressor of MEK null proteins; affects the transcription of DAF-16 target genes	1.64
	emb-5	regulator of transcriptional elongation by RNA polymerase II	1.71
	nono-1	conserved nuclear protein, forms a complex with the mRNA export factor NXF-1	1.85
	ceh-38	homeobox protein, DNA-binding regulatory protein	1.92
	hmg-1.2	positive regulation of transcription from RNA polymerase II promoter	2.24
	sta-1	signal transducer and activator of transcription 1	2.46
	hcf-1	transcriptional regulator that associates with histone modification enzymes	2.5
	rtfo-1	RNA polymerase-associated protein, component of the PAF1 complex	2.56
	spt-4	transcription elongation factor	2.63
Synthetic multivulva class B	lin-53	synthetic multivulva class B (synMuvB) protein, transcription factor member of the (DRM) complex	1.78
	lin-37	synthetic multivulva class B (synMuvB) protein, transcription factor member of the (DRM) complex	1.89
	lin-35	synthetic multivulva class B (synMuvB) protein, transcription factor member of the (DRM) complex	2.83
	dpl-1	synthetic multivulva class B (synMuvB) protein, transcription factor member of the (DRM) complex	3.02
mRNA processing	teg-4	pre-mRNA splicing factor, tumorous enhancer of Glp-1	1.51
	uaf-2	splicing factor	1.76
	pap-1	poly(A) polymerase	1.83
	rnp-4	core component of the splicing-dependent multiprotein exon junction complex (EJC)	1.83
	prp-65	pre-mRNA processing factor 6	1.95
	rsp-4	splicing factor	2.33
	lsm-7	mRNA splicing factor, via spliceosome	2.9
Ribonucleoproteins (RNP)	snr-3/-6/-7	heptameric complex required for biogenesis and function of the snRNPs	1.88/2.62/2.91
	fust-1	FUS/TLS RNA binding protein homolog	1.95
	rnp-2	small nuclear ribonucleoprotein (snRNP)-associated protein RNP-2/U1A	2.48
	rop-1	protein component of the Ro ribonucleoprotein (RNP) complex	4.06
	hrpf-1/-2	orthologous to human hnRNP F and hnRNP H, act as pre-mRNA splicing factors	4.12/2.71
Transport	pgl-1/-3	P granule abnormality protein	1.52/3.72
	npp-2/4-4/-7/-9	nuclear pore complex proteins	1.57–4.04
	xpo-1	nuclear export receptor	1.61
	ran-1/2	GTP-binding nuclear protein	1.64/1.63
	hel-1	spliceosome RNA helicase DDX39B homolog	1.79
	thoc-3	THO complex (transcription factor/nuclear export) subunit	1.94
	iff-1	eukaryotic translation initiation factor 5A-1	2.01
	nxf-1	nuclear RNA export factor 1	2.14
	nxt-1	NTF2-related export protein	2.15
	imb-1	importin beta family	2.15
	aly-1/-3	Ref/ALY RNA export adaptor family	3.45/1.58
Extracellular
Transthyretins	ttr-5/-6/-15/-17	transthyretin-like protein	1.8–4.17
FA binding proteins/transporters	nrf-5	lipid-binding protein	1.9
	lbp-1	FA-binding protein	2.5
Others	mec-5	collagen unique in the number of Gly-X-Y repeats	1.58
	egl-3	prohormone convertase	1.92
	sod-4	extracellular superoxide dismutase (Cu-Zn)	8.32
Plasma Membrane
Transmembrane channel proteins	inx-3/-6/-12/-16	innexin	1.7–3.26
Heterotrimeric G proteins	gpb-1	guanine nucleotide-binding protein subunit beta-1	1.65
	goa-1	heterotrimeric G protein alpha subunit Go (Go/Gi class)	1.93
	egl-30	heterotrimeric G protein alpha subunit Gq (Gq/G11 class)	1.95
	eat-16	RGS protein, interacts with the egl-30 and goa-1 signaling pathways	2.42
ATPases	eat-6	alpha subunit of a sodium/potassium ATPase	1.7
	nkb-1	sodium/potassium-transporting ATPase subunit beta-1	1.79
	vha-5	V-type proton ATPase subunit a	1.97
	mca-3	calcium-transporting ATPase	3.32
	catp-3	cation transporting ATPase	3.46
Amino acid, ion, and ATP transporters	mrp-2/-7	ATP-binding cassette transporter, member of the ABCC subfamily	1.51/1.58
	abts-1/3	sodium-driven chloride-bicarbonate transporter	1.77/1.75
	atgp-1/-2	amino acid transporter glycoprotein subunit	2.16/1.51
	haf-2/-7	transmembrane protein of the ATP-binding cassette transporter superfamily	2.75/2.16
Endocytosis andvesicle trafficking	aex-3	MAP kinase protein required for intracellular vesicle trafficking as well as synaptic vesicle release	1.83
	arf-6	ADP-ribosylation factor	1.83
	itsn-1	endocytic adaptor protein to regulate cargo sorting through the endolysosomal system	1.84
	rab-3	involved in exocytosis by regulating a late step in synaptic vesicle fusion	1.9
	snap-29	SNARE, soluble essential protein for fusion of cellular membrane	2.12
	dyn-1	dynamin GTPase, its activity is required for endocytosis, synaptic vesicle recycling	2.13
	arl-8	Arf-like small GTPase, regulates transport of axonal presynaptic vesicle protein cargo	3.05
	sqst-1	ATP-binding cassette transporter, member of the ABCC subfamily	30.27
Proteins Significantly (FDR < 5%) Decreased in Abundance in UV-Treated versus Untreated xpc-1;csb-1 Double Mutants
Ribosomes
Large subunit	rpl-4/-5/-6/-7/1(6/-0S	60S ribosomal proteins	0.42–0070
Small subunit	ubl-1	ubiquitin-like protein 1-40S ribosomal protein	0.4
	rps-1/-4/-6/-8/-	40S ribosomal proteins	0.44–0.69
Translation initiation factors	iffb-1	eukaryotic translation initiation factor eIF5B	0.56
	eif-3.D	eukaryotic translation initiation factor 3 subunit D	0.62
	eif-3.E	eukaryotic translation initiation factor 3 subunit E	0.65
	egl-45	eukaryotic translation initiation factor 3 subunit A	0.66
Others	rrbs-1	ribosome biogenesis regulatory protein homolog	0.31
	mrps-2/-5/-22/-30	mitochondrial ribosomal protein, small	0.46–0.69
	mrpl-15/-22/-35/-38/-40/-50	mitochondrial ribosomal protein, large	0.49–0.69
UPS Machinery and Chaperones
Ubiquitous proteasome system	sao-1	suppressor of aph-1,regulates the notch receptor signaling pathway	0.26
	ubc-26	ubiquitin conjugating enzyme	0.31
	hecd-1	E3 ubiquitin protein ligase 1 homolog, involved in ubiquitin-dependent protein catabolic process	0.50
	cul-3	RING-finger protein, form the catalytic core of an SCF-type E3-ubiquitin ligase complex	0.58
	rpt-4	ATPase subunit of the 19S regulatory complex of the proteasome	0.68
Chaperones	dnj-2/-11/-13/-27/-29	ribosome-associated molecular chaperones	0.09–0.69
	D2030.2	orthologous to human atP-dependent Clp protease atP-binding subunit clpX-like, hsp100 family	0.48
	cct-5/-6	T-complex protein 1 subunit epsilon and zeta	0.67–0.67
Mitochondria
Mitochondria machinery	mspn-1	mitochondrial sorting homolog	0.42
	timm-23	translocase, inner mitochondrial membrane	0.47
	W02B12.9	mitochondrial iron transporter that mediates iron uptake enzymes	0.49
	K11H3.3	putative tricarboxylate transport protein, mitochondrial	0.5
	acdh-13	acyl-coenzyme A (CoA) dehydrogenase involved in FA beta-oxidation	0.53
	F53F10.3	mitochondrial pyruvate carrier 2	0.58
Peroxisomes
Peroxisome machinery	acox-1	acyl-coenzyme A oxidase	0.3
	C48B4.1	peroxisomal acyl-coenzyme A oxidase 5	0.34
	gstk-15	glutathione S-transferase kappa 1	0.35
	daf-22	ortholog of human sterol carrier protein SCP2, which catalyzes the final step in peroxisomal fatty acid beta-oxidation	0.6
	ctl-2	peroxisomal catalase 1	0.7
Endoplasmic Reticulum
ER chaperones	hsp-3	heat shock 70 kDa protein C	0.6
	phy-2	prolyl 4-hydroxylase subunit alpha-2	0.11
Others	dpy-185	prolyl 4-hydroxylase subunit alpha-1	0.11
	C14B9.2	protein disulfide-isomerase A4	0.28
	enpl-1	endoplasmin homolog	0.58
	Fkb-3/-4/-5	peptidyl-prolyl cis-trans isomerase	0.35/0.32/0.20
	srpa-68	signal recognition particle subunit SRP68, has 7S RNA binding activity	0.45
	pdi-2	protein disulfide-isomerase 2	0.55
FA Metabolic Processes
Fatty acid biosynthesis	5/fat-1/-2/-4/-6	omega-3 FA desaturases	0.02/0.24
	D1/acs-1/-5/-7/-16/	FA CoA synthetase family	0.06–0.57
	fasn-1	FA synthase	0.20
	pod-2	acetyl-CoA carboxylase, catalyzes the first step in de novo FA biosynthesis	0.30
	elo-1/-2/-5/-6	elongation of very long chain FA proteins	0.31/0.45
	ech-6/-7	enoyl-CoA hydratase	0.45/0.31
	bcat-1	branched-chain-amino-acid aminotransferase	0.53
Glycerolipid and glycerophospholipid metabolism	sams-1	S-adenosylmethionine synthase 1	0.42
	mboa-3	membrane-bound O-acyl transferase	0.48
	acl-6/-7	glycerol-3-phosphate acyltransferase predicted to play a role in triacylglycerol biosynthesis	0.51/0.32
	ckb-4	choline/ethanolamine kinase	0.51
	ckc-1	choline/ethanolamine kinase	0.55
SL metabolism	sptl-2/-3	glycerol-3-phosphate acyltransferase, plays a role in triacylglycerol biosynthesis	0.25–0.28

Nuclear Import and Export Transport Is Enhanced upon DNA Damage

Increased proteins regulating translation, spliceosome assembly, and nuclear-cytoplasmic transport suggest an involvement of RNA biogenesis and translocation in the DNA damage response (Table 1). The nuclear pore complex proteins (nucleoporins [NPPs]), together with Ran-GTPases, play an important role not only in nuclear import and export and nuclear envelope (NE) assembly dynamics but also in the localization of MEL-28 (Fernandez and Piano, 2006), a structural NE component that regulates the distribution of the integral nuclear-envelope proteins EMR-1, LMN-1, LEM-2, and BAF-1 (Table 1) (Galy et al., 2006). These nuclear proteins provide an anchor attaching chromosomes to the nuclear membrane, are required for proper chromosome segregation (Liu et al., 2003), and promote the reorganization of damaged chromatin upon UVC and ionizing radiation (IR)-induced DNA damage (Dittrich et al., 2012). When responding to stress, BAF-1 is immobilized at the nuclear lamina, stabilizes the chromatin structure, and influences the gene expression via histone post-translational modifications (Montes de Oca et al., 2011). Exposure of human cells to UV-induced DNA damage causes BAF-1 to dynamically interact with the histone H3/H4 ubiquitin ligase complex (CUL4-DDB-ROC1), facilitating the recruitment of repair proteins to the damaged DNA (Montes de Oca et al., 2009). BAF-1 expression is regulated by transcription factors that modulate lifespan, including SKN-1, PHA-4, DAF-16, and ELT-3 (Bar et al., 2014).

Similar to aged IIS mutant worms (Halaschek-Wiener et al., 2005) and cells responding to DNA damage (Matsuoka et al., 2007), proteins belonging to the nuclear category implicated in DNA replication and cell-cycle progression (CDK-1, MCM-2, MCM-7, and RFC-4) were decreased in abundance upon UV treatment (Table S1).

Differences in Ion Transport and Synaptic Transmission in Worms Treated with UV Irradiation

We also observed upregulation of proteins belonging to plasma membrane and extracellular space, suggesting possible intra- or extracellular trafficking of signals from genotoxically compromised cells (Figures 1B and 1D). The plasma membrane category contains transmembrane channel proteins, ATPases, amino acid, ion and ATP transporters, and heterotrimeric G proteins (key regulators of G protein-coupled receptor [GPCR] signaling) (Table 1). GPCR signaling has been implicated in fundamental aspects of development and behavior, including the synaptic transmission in the ventral cord motor neurons (Nurrish et al., 1999). Excitable cells display the highest expression of heterotrimeric G proteins together with components of the endocytic pathway involved in the initial vesicles assembly (ARF-6, ARL-8, and DYN-1), vesicle fusion (SNAP-29 and AEX-3), and vesicles recycling through the endo-lysosomal system (ITSN-1 and SQST-1). The upregulation of those proteins might indicate neuronal signals responding to DNA damage (Table 1). DNA repair defects have been linked to the impaired neuronal development in various human congenital progeroid syndromes, including CS, and to age-related neurodegenerative disorders such as Alzheimer’s disease (AD) and amyotrophic lateral sclerosis (ALS) (Martin, 2008). Consistent with neuronal developmental processes being affected by unrepaired DNA damage, we found also elevated levels of proteins implicated in axonal outgrowth (EAT-6, CAM-1, UNC-44, and TBB-4) and neuronal positioning during development (SAX-7, WRK-1, UNC-33, and UNC-37) (Table S1).

Increased extracellular proteins were mainly hormone carrier transthyretin (TTR)-related factors, also reported as elevated in aged C. elegans (Liang et al., 2014) and associated with neuroprotection in a murine AD model (Buxbaum et al., 2008). The extracellular Cu²⁺/Zn²⁺ superoxide dismutase SOD-4 and some lipid binding proteins and transporters (NRF-5, LBP-1, and EGL-3), which sequester respectively potentially toxic peroxidation products and toxic fatty acids (FAs), were also upregulated (Table 1).

Downregulated Protein Clusters upon Genotoxic Stress

A large number of ribosomal proteins, including components of the small (40S), large (60S), and mitochondrial ribosome subunits, together with components of the translation machinery, were downregulated upon UV treatment (Table 1). A similar drop was observed also for factors involved in protein homeostasis, localized between cytoplasm, mitochondria, endoplasmic reticulum (ER), peroxisomes, and for regulators of FA metabolism (Figures 1B and 1D; Table 1). This general decline in protein synthesis and dampening of metabolic processes upon UV treatment is consistent with previous reports from proteomic studies of aged animals (Ben-Zvi et al., 2009, Narayan et al., 2016), supporting parallels between the DNA damage response and aging.

Protein Targeting for Degradation

Impaired protein homeostasis has been suggested as a hallmark of aging (Powers et al., 2009). Protein-fold stabilization restores the structure of misfolded polypeptides, or remove and degrade, via the proteasome or the lysosome, aberrant proteins. Upon UV treatment, many components of the proteostasis network, as chaperones, ubiquitin ligases, and members of the ubiquitin-proteasome system (UPS) machinery, together with ER, peroxisomal and mitochondrial homeostasis-related proteins were downregulated (Table 1). This included the E3 ubiquitin ligase Y54E10A.11 that contains a RING-type-domain homologous to human TRAIP, implicated in the cellular UV response. Missense mutations in the TRAIP RING-domain have been identified in patients with premature aging syndromes (Harley et al., 2016). Y54E10A.11 is a component of the ribosome quality-control complex (RQC), which recognizes stalled ribosome and associates with the 60S subunit, allowing the ubiquitination and extraction of incompletely synthesized nascent polypeptides (Defenouillère et al., 2013). The translation stress specifically sensed by the RQC is communicated to the transcription factor HSF-1 (Brandman et al., 2012), which in turn promotes lifespan extension (Morley and Morimoto, 2004), suggesting a combined strategy to play a role in both longevity and stress responses.

Impairment of chaperones and UPS machinery give rise to misfolded proteins that that are imported into lysosomes during chaperone-mediated autophagy, or sequestered in autophagosomes during macroautophagy (Megalou and Tavernarakis, 2009). Upon DNA damage, we found an upregulation of macroautophagy sub-pathway members: ATG-3, ATG-18, and SQST-1, the p62 homologous. Autophagy, via the elimination of SQST-1, was recently implicated in the regulation of the DNA damage response via chromatin ubiquitination (Wang et al., 2016). The decrease in protein synthesis, together with the impairment of protein refolding and degradation mechanisms and the decreased mitochondrial homeostasis, suggests general organismal energy depletion upon UV-induced DNA damage. Given the role of autophagy in degrading aberrant proteins and in recycling nutrients and energy, we hypothesized a proteostatic shift toward autophagy to allow the organism tolerating the consequences of impaired protein homeostasis (Figure 2A).

Figure 2.

Proteostatic Shift Leads to Induction of Autophagy, which Is Required for UV Resistance

(A) Proteostatic shift from protein synthesis and degradation mechanisms toward autophagy.

(B) Immunoblotting of the autophagy marker LGG-1(I) and LGG-1(II)::GFP. LGG-1 becomes lipidated after UV-induced DNA damage and starvation (representative of three independent experiment shown).

(C) WT, atg-3(bp412), and atg-9(bp564) L1 larvae were irradiated or mock treated, and developmental stages were evaluated 48 hr later. An average of three independent experiments per strain and dose is shown; >15 individuals were analyzed per experiment. Error bars denote standard deviation. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 (two-tailed t test compared with WT).

To monitor autophagy, we used a GFP-fusion transgene of the ubiquitin-like, microtubule-associated Atg8/LC3 ortholog LGG-1 required for autophagic vesicle growth (Levine and Klionsky, 2004). Within 4–10 hr after UV treatment, we observed significantly increase of the lipidated LGG-1(II) form indicative of autophagy (Figures 2B and S3). To assess whether autophagy was required for withstanding DNA damage, we tested the UV sensitivity of two autophagy mutants, atg-3(bp412) and atg-9(bp564). We observed a significantly higher sensitivity of the autophagy mutants than of wild-type (WT) worms (Figures 2C and S4), suggesting that proteins involved in the formation of autophagosomes are essential for enduring DNA damage. The impaired UPS machinery and the increased autophagy activity are discussed below in the network analysis (Figure 5).

Figure 5.

Network of Interactions between Proteins that Are Significantly Regulated in xpc-1;csb-1 Double Mutants upon UV Treatment

FDR < 5%. Dark blue, >1.5-fold downregulated; dark red, >1.5-fold upregulated. Different shapes indicate the significantly changed phosphosites normalized to the proteome.

Metabolic Alterations in Worms Treated with UV Irradiation

Translation and autophagy are regulated, in parallel to IIS signaling, by the target of rapamycin (CeTOR) LET-363 in complex with the raptor protein DAF-15 to influence cell growth and longevity (Wullschleger et al., 2006). Upon UV treatment, autophagy-related proteins as well as CeTOR and IIS pathway members were elevated (Table S1), reminiscent of the increase of the same members of those pathways during aging (Narayan et al., 2016). Autophagy has been reported to mobilize lipids via the breakdown of lipid droplets (lipophagy) (Singh et al., 2009). Upon UV treatment, we observed a decrease of proteins involved in lipid metabolism and localized between the cytoplasm, ER, peroxisomes and mitochondria (Table 1), similar to their decrease during C. elegans aging (Narayan et al., 2016).

Taken together, these observations suggest that amid persistent DNA damage worms reduce DNA replication and translation, thus potentially avoiding the production of aberrant proteins. Moreover, protein-refolding mechanisms are reduced, whereas autophagy is elevated, suggesting a rerouting of protein recycling as part of metabolic shift in response to the DNA damage.

Correlation between Proteome and Transcriptome

To address the role of transcriptional responses to UV-induced DNA damage, we compared the proteomes with previously published transcriptome data of xpa-1 mutants as NER deficient as the xpc-1;csb-1 mutants (and phenotypically identical in response to UV irradiation) (Mueller et al., 2014). We found a significant moderate positive correlation (r = 0.347) between the significantly changed transcripts and proteins upon UV treatment (Figure 3A; Table S3), suggesting that the expression of only a part of proteins can be explained by transcription, while a large fraction is subject to post-transcriptional regulation (see later discussion).

Figure 3.

Response to Persistent DNA Damage Correlates with Starvation Stress and Aging Proteomes

(A–C) Correlation analyses (A) between proteome of xpc-1;csb-1 double mutants (FDR < 5%) and transcriptome of xpa-1 mutants after UV treatment (similarly regulated proteins and genes in red and green; specific protein clusters are detailed in Table S3), (B) between proteins detected in xpc-1;csb-1 double mutants upon UV treatment versus aging in WT worms (p < 2.2 × 10⁻¹⁶ for the three Pearson correlation coefficients, r), and (C) between proteins changed in abundance of at least 2-fold (FDR < 5%) in xpc-1;csb-1 double mutants upon UV treatment versus starvation.

Correlation between Proteome upon UV Treatment and Aging

To address whether proteome changes in response to DNA damage might bear similarities to those occurring during aging, we conducted a correlation analysis between proteomes of UV-treated xpc-1;csb-1 double mutants, unable to repair the UV-induced DNA damage, and WT worms during aging (Walther et al., 2015). Indeed, the proteomes of UV-treated NER deficient animals and WT worms during aging were positively correlated (day 12, r = 0.26; day 27, r = 0.34; Figure 3B), suggesting that the regulation at the protein level upon persistent DNA damage bears similarities to proteome alterations during aging. The significant positive correlations were more striking when we compared the DNA damage responses of L1 larvae with those of aging adult animals. The similarly regulated processes revealed a general enrichment of factors involved in FA metabolism, oxidative stress response, UPR, and IIS.

Correlation between Proteome upon UV and Starvation Treatment

L1 larvae arrest their growth not only upon genotoxic treatment but also for extended periods of time in the absence of food and resume developmental growth only when food becomes available. We have previously found similar and contrasting transcription responses between starvation conditions and UV-induced DNA damage (Mueller et al., 2014). In parallel to UV treatment, we also performed starvation experiment in xpc-1;csb-1 double mutants: three independent biological replicates were analyzed, with excellent reproducibility (r > 0.95 for biological replicates) (see also Figure S1). We obtained a positive Pearson correlation between the proteomes of UV-irradiated and starved animals (r = 0.77) (Figure 3C). The similarities were composed of proteins associated with chromatin, vesicle/neurotransmitter trafficking including heterotrimeric G proteins implicated in the starvation-induced activation of the Ras-MAP kinase pathway (You et al., 2006), and metabolic pathways involved in the synthesis and use of carbohydrate, amino acid, and lipids (Tables 1 and S1). Key enzymes involved in FA biosynthesis (Figure 4A) and playing important roles in FA accumulation and consumption during lifespan (Horikawa et al., 2008) were downregulated (Tables 1 and S1). The expression of the same class of genes related to lipid metabolism has been found significantly decreased in the UV-irradiated and photoaged human skin, suggesting that inhibition of de novo lipid synthesis could have a detrimental effect, leading also to collagen destruction (Kim et al., 2010) (Table S1).

Figure 4.

Attentuated FA Synthesis and Alterations in Lipid Classes

(A) Key members of the FA biosynthetic coupled to SL and phospholipid metabolic pathways were significantly decreased in abundance in xpc-1;csb-1 double mutants upon starvation and UV treatment.

(B) TLC detection and quantification (histogram) of FAs and triacylglycerols in xpc-1;csb-1 double mutants upon UV and starvation.

(C and D) Changes in the amount of three SLs subclasses (ceramides, sphingomyelins, and glucosylceramides) (C) and of the five glycerophospholipids subclasses (phosphatidylcholine [PC], phosphatidylethanolamine [PE], PI, phosphatidylserine [PS], and phosphatidylglycerol [PG]) (D) assessed by MS analysis.

Significant levels of pairwise comparisons are indicated by p values: ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001.

Lipidomics Analysis upon DNA Damage and Starvation

Prompted by the alterations in FA biosynthesis enzymes, we next traced lipid profiles of xpc-1;csb-1 double mutants upon UV treatment and starvation by using thin-layer chromatography (TLC) and MS. We observed a decrease in triacylglycerols, the storage form of FAs (Figures 4B and S4) that is consistent with the worms’ deriving energy from degradation of fat stored to survive stress such as food deprivation (Elle et al., 2012).

The downstream products of these FA biosynthetic pathways are normally used to synthesize more complex lipids: saturated FAs (SFAs) serve as building blocks for the sphingolipids (SLs), whereas both SFAs and unsaturated FAs (UFAs) are incorporated into glycerophospholipids (Figure 4A). SLs are highly conserved components of cell membranes having regulatory roles in growth control and aging in a wide range of organisms (Cutler et al., 2014). SL works as an intermediate for the production of ceramide (Cer), a key product for the synthesis of glucosylceramide and sphingomyelin (SM) (Zhang et al., 2015) (Figure 4A). Cer is produced from SM in UV- and IR-treated mammalian cells (Zeidan et al., 2008), whereas an increased synthesis of SM from Cer is associated to accelerated development and aging (Cutler et al., 2014). Similarly to aging studies, MS-based quantitative SL profiling showed a general increase in SM and decrease in Cer upon both treatments (Figures 4C and S4), potentially as a consequence of the impaired SFAs biosynthesis. elo-5 mutants, deficient for monomethyl branched chain FA (mmBCFA) synthesis, arrest development similar to starved L1 larvae (Kniazeva et al., 2008) and could be rescued by SFA-derived SLs, d17iso-glucosylceramides (d17iso-GlcCer), together with downstream factors of the CeTOR pathway (Zhu et al., 2013). Intriguingly, upon UV treatment, we observed an increased abundance of members of the CeTOR pathway (Table S1) and of d17iso-GlcCer (Figures 4C and S4). In line with a previous study (Kniazeva et al., 2008) reporting stable mmBCFA levels in starved L1 larvae, we also observed stable levels of d17iso-GlcCer species upon starvation (Figures 4C and S4). The role of the GlcCer/TOR pathway in promoting development independently from the IIS and DAF-7/TGFβ-signaling (Zhu et al., 2013) suggests that it regulates the developmental response to UV-induced DNA damage.

Another major component of cellular membranes is the lipid class of glycerophospholipids, synthesized from the intermediate phosphatidic acid, through a series of reduction and acylation reactions (Figure 4A). Phosphatidic acid is dephosphorylated to yield DAG, which is converted into phosphatidylcholine (PC) and phosphatidylethanolamine (PE), which can be both intermediates for the formation of phosphatidylserine (PS). PS and phosphatidylinositol (PI) are generally synthesized from cytidine diphosphatediacylglycerol (CDP-DAG), substrate for the synthesis of phosphatidylglycerol (PG) and cardiolipin (Zhang et al., 2015). Quantitative glycerophospholipids MS profiling, upon starvation and UV treatment, showed a change of the DAG downstream products, indicating a preferential direction in the phospholipid synthesis (Figures 4D and S4). Upon UV, the PC and the PC-derived PS were increased, whereas PE was reduced. In contrast, upon starvation, the PE and the PE-derived PS were elevated, whereas PC was decreased. Other CDP-DAG-derived phospholipids (PI and PG) were not changed, except for a significant reduction of PG in response to starvation (Figures 4D and S4). Taken together, these observations suggest that the worms respond to persistent DNA damage by a metabolic shift reminiscent of adaptations during starvation (Larance et al., 2015) and aging (Liang et al., 2014).

Proteome and Phosphoproteome-Coupled Analysis to Build a Regulatory Network

In order to also follow the dynamics of PTMs, we extended our label-free quantitative MS analysis by performing phosphopeptide enrichment with the titanium bead (TiO₂) method. The correlation plot of the phosphoproteome dataset upon each treatment (untreated, UV treated, and starvation) shows how the biological replicates cluster together, in a correlation range from 0.7 to 1, as reported in the color key map (Figure S5A). The distributions of the individual phosphorylated residues (Ser/Thr/Tyr) (Figure S5B) and the number of phospho-groups per peptide we detected were similar to those obtained in previous studies (Lundby et al., 2012). Among the 7,430 detected phosphosites, we identified 3,276 significantly modulated in response to UV treatment, with 1,571 more than 1.5-fold downregulated and 1,705 more than 1.5-fold upregulated (p < 0.05) (Table S4). We used Cytoscape and the C. elegans data repository (WormBase) interaction data to generate a protein-protein interaction map (Cytoscape), including only significantly regulated proteins and phosphosites normalized to the proteome (Figure 5). The central node of the network is the DAF-2 protein, a component of IIS signaling, a pathway that has been implicated in the regulation of both the DNA damage response and longevity (McElwee et al., 2004, Mueller et al., 2014). The main clusters of upregulated proteins arising from the DAF-2 central node encompass chromatin organizers, the synthetic multivulva class B family of proteins, the CeTOR, and proteins involved in nuclear-cytoplasmic transport (Figure 5; Table 1). Nuclear transport proteins as PGL-1 and PGL-3 are intermediary nodes between DAF-2 and some autophagy proteins, in particular with the highly upregulated SQST-1, which together with other upregulated components of the endocytic pathway is involved in the neuronal synaptic machinery (Table 1). The upregulated synaptic machinery for the hormone and neurotransmitter release, like heterotrimeric G proteins, or for the mechanosensation, like the MEC proteins, were indirectly linked to DAF-2. As mentioned earlier, factors belonging to the UPS machinery, as well as some chaperones, and members of the ER proteostasis network were downregulated (Figure 5; Table 1). FA metabolic enzymes (Table 1) and proteins involved in amino-acid biosynthesis (SAMS-1), development (DAO-2, DNJ-25, CALU-1, and CUA-1), and stress response (NSY-1 and LYS-7) were also decreased. Using the BiNGO tool, we determined a significant overrepresentation of the GO biological processes larval development, cellular biosynthesis modulating translation, and organic acid biosynthesis, in particular the FA biosynthetic processes within this interaction map (Figure S6).

We next derived signaling pathways that respond to persistent DNA damage (Figure 6A) on the basis of the combining proteins that showed significant alterations at the proteome and/or phosphoproteome level. As central signaling platform appears EGF signaling, which has been linked to development, metabolism, and longevity in C. elegans (Iwasa et al., 2010). The EGF signaling cascade is transduced through the phospholipase Cγ (PLC)/protein kinase C (PKC), the PI 3-kinase (PI3)/AKT, and the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathways (Jorissen et al., 2003) as well as through the RAS/extracellular signal-regulated kinase (ERK) to regulate protein homeostasis via the expression of antioxidant genes and the stimulation of the UPS activity via the activation of SKR-5 protein (Liu et al., 2011). EGF signaling also regulates cell growth and survival via the PI3K/AKT kinase cascade that impacts the activity of CeTOR and the IIS effector DAF-16 (Hay, 2011).

Figure 6.

Network Analysis of Proteome and Phosphoproteome Alterations in Response to Persistent DNA Damage

(A) Network of interactions between proteins regulated in xpc-1;csb-1 double mutants upon UV treatment. Symbols are as follows: full circles, proteins detected by MS as downregulated (blue) or upregulated (red) or not significantly regulated (white); dotted circles, proteins that are not quantified by MS; and stars, phosphopeptides detected by MS as decreased (blue) or increased (red). p < 0.05.

(B) WT, aak-2(gt33), and aak-2(ok524) L1 larvae were irradiated or mock treated, and developmental stages were evaluated 48 hr later. An average of three independent experiments per strain and dose is shown; >540 individuals were analyzed per experiment. Error bars show standard deviation; ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗ p < 0.001 (two-tailed t test compared with WT).

EGF and G protein signaling regulate the PLC-mediated hydrolysis of PI 4,5 bisphosphate (PIP2) into the second messengers diacylglycerol (DAG) and inositol 1,4,5-triphosphate (IP3) (Rhee, 2001). Although IP3 promotes calcium release (Ca²⁺), DAG is an intermediate of the glycerophospholipids synthesis (Figure 4) and functions as a cofactor for the activation of PKC. The cluster of proteins downstream of PKC have been implicated in the regulation of daf-2 IIS-dependent control of dauer formation (Monje et al., 2011) and in the secretion of synaptic vesicles at motor neurons (Sieburth et al., 2007). Consistently, we detected a high abundance of the downstream components of the G protein signaling, involved in the neuronal synaptic machinery, and mediating the initial vesicles assembly (Figures 1D, 5, and 6A; Table 1). The heterotrimeric G proteins α subunits, EGL-30 and GOA-1, not only mediate serotonin signaling, promoting intracellular vesicle trafficking and synaptic transmission (Nurrish et al., 1999) but also regulate the expression of DAF-7, a member of the TGFβ-signaling pathway that during larval development regulates DAF-16 and STA-1 nuclear localization (Myers, 2012, Shaw et al., 2007, Wang and Levy, 2006). The activity of the two G protein subunits EGL-30 and GOA-1 in regulating neurotransmitter secretion is itself regulated by the guanine nucleotide exchange factor RIC-8 (Miller et al., 2000). RIC-8 is able to activate another α subunit of the heterotrimeric G proteins pool, GSA-1, which in turn activates the adenylyl cyclase ACY-1 to produce cyclic AMP (cAMP). This signaling cascade leads to the activation of the regulatory subunits (KIN-1 and KIN-2) of cAMP-dependent protein kinase A (PKA), to modulate growth and locomotion (Schade et al., 2005). Once activated, PKA acts on the cAMP-responsive element (CRE)-binding protein (CREB; CRH-1 in C. elegans), modifying its phosphorylation status and thus altering its subcellular localization. This stimulates the association of CRH-1 with its cAMP-regulated transcriptional co-activator (CRTC; CRTC-1 in C. elegans). Together these two factors target CREs on promoter genes, which regulate the glucose and lipid metabolism (Altarejos and Montminy, 2011). The downregulation of CRH-1 and CRTC-1 in an AMP-activated protein kinase (AMPK; AAK-2 in C. elegans)- and calcineurin-dependent manner, induces transcriptional responses that modulates longevity (Mair et al., 2011). Given that AAK-2 plays a central role in controlling energy metabolism and regulating longevity through the CeTOR and the daf-2-mediated IIS pathways (Curtis et al., 2006), we wondered whether the AMPK homolog might be involved in the response to persistent DNA damage. Indeed, two independent aak-2 mutant alleles showed a significantly more sensitivity to UV treatment than WT worms (Figures 6B and S4). Another factor involved in the determination of adult lifespan via the negative regulation of the IIS downstream target, DAF-16, is the enzyme HCF-1 (Li et al., 2008). This factor works also as transcriptional regulator of chromatin modification and histone phosphorylation (Lee et al., 2007). Consistently, we also found a number of histone modifiers and chromatin organizers highly upregulated in response to UV treatment (Figures 1D and 6A; Table 1). Other increased proteins belong to the cohesion complexes that require the activity of the four upregulated proteins SMC-1, SMC-3, COH-1, and SCC-3 for chromosome segregation and the repair of double-strand breaks (Baudrimont et al., 2011) (Table 1). AIR-1 and PLK-1 kinases or the cleavage by separase (SEP-1) are required for the dissociation of this cohesion complex allowing the segregation of sister chromatids during mitosis (Tsou et al., 2009). Taken together, the network analysis reveals an intricate network of differentially regulated signaling nodes and centrally places IIS, EGF, and AMPK signaling in the DNA damage response.

Discussion

The similarities between the proteome changes we observe upon acute DNA damage and those occurring during aging are consistent with the causal role of DNA damage accumulation in the physiological adaptations in aged animals. Previous experiments based primarily on transcriptome analyses of NER-deficient mice suggested an adaptive “survival response” to accumulating DNA damage during aging that preserves tissue functionality by attenuating the somatic growth axis (Garinis et al., 2009, Niedernhofer et al., 2006, van der Pluijm et al., 2007). The proteomics, phosphoproteomics, and lipidomics alterations we observed in response to persistent DNA damage further support such a shift of the organism’s resources to preservation of somatic functioning. Our analysis places DAF-2 as a central hub, consistent with the role of the IIS effector transcription factor DAF-16 in counteracting the detrimental consequences of DNA damage (Mueller et al., 2014). Similarly to previous studies of aged IIS mutant worms (Halaschek-Wiener et al., 2005) and cells challenged with DNA damage (Matsuoka et al., 2007), we detected a reduction in DNA replication-associated processes and induction of proteins involved in nuclear mechanisms, consistently with chromatin remodeling modulating repair, replication, and transcription upon DNA damage. As in proteome studies of aged animals (Liang et al., 2014), levels of ribosomal proteins and the translation machinery were reduced potentially to avoid the production and accumulation of aberrant proteins. The proteostatic pathways were shifted toward autophagy, which might serve as a compensatory response when protein homeostasis is impaired and which we find to be required for the animals to withstand DNA damage.

We observed many similarities between the proteome alterations upon DNA damage and those during aging and starvation. We detected significant changes of metabolic pathways involved in the synthesis and use of carbohydrate, amino acid, and lipids upon persistent DNA damage. Many enzymes involved in FA biosynthesis showed significant decreases in abundance, similarly to the reports in proteomic studies of IIS-deficient worms (Depuydt et al., 2014), upon starvation (Larance et al., 2015), and during aging (Narayan et al., 2016). Our quantitative lipid profiling of worms undergoing persistent DNA damage showed a dampening also of fat biosynthesis and a differential regulation of the complex downstream targets. Therefore, it will be interesting to investigate the role of regulators of lipid metabolism in the DNA damage response.

The network analysis combining proteome and phosphoproteome datasets reveals an intricate connection of differentially regulated signaling nodes and assigns central roles for the IIS regulator DAF-2 and the EGF- and AMPK-like signaling pathways in response to DNA damage. Mutations affecting the AMP-activated protein kinase AAK-2, playing a central role in controlling energy metabolism and regulating longevity through the CeTOR and IIS pathways (Curtis et al., 2006), lead to an increased to DNA damage sensitivity. Members both of the CeTOR and IIS pathways were upregulated upon UV treatment, reminiscent of recent C. elegans studies during aging (Narayan et al., 2016), reinforcing the concept of response to accumulating DNA damage during the natural aging process.

In conclusion, our analysis of the proteome, lipidome, and phosphoproteome after UV treatment in NER-deficient C. elegans provides a comprehensive picture of the response processes to persistent DNA damage in a metazoan animal model. The identified proteins and pathways open up previously unexplored avenues for establishing a complete model of organismal responses to persistent DNA damage. This complex in vivo analysis provides a starting point to translate the insight into the signaling networks involved in an animal’s response to UV treatment to higher organisms and ultimately to humans, thus helping decipher outcomes of NER deficiency syndromes and gaining a better understanding of the consequences of DNA damage in the aging process.

Experimental Procedures

C. elegans Strains

Strains were maintained at 20°C on nematode growth medium (NGM) with E. coli strain OP50. N2 (Bristol; WT), TG38 aak-2(gt33), RB754 aak-2(ok524), BJS724 atg-3(bp412), BJS725 atg-9(bp564), BJS21 xpc-1(tm3886);csb-1(ok2335), BJS173 xpc-1(tm3886);csb-1(ok2335);daf-16(mu86);zIs356[daf-16::GFP; rol-6(su1006)], and DA2123 (adIs2122 [lgg-1p::GFP::lgg-1; rol-6(su1006)]).

Proteome

xpc-1;csb-1 double mutants were synchronized by bleaching and fed for 3 hr or left under starvation. Six hours after UV (310 nm, Phillips UV6, Waldmann UV236B) or mock treatment, worms were collected in extraction buffer (50 mM HEPES KOH [pH 7.5], 300 mM NaCl, 1 mM EDTA, 1% [v/v] Triton X, 0.1% [w/v] sodium deoxycholate, 10% [v/v] glycerol, and complete protease inhibitor cocktail; Roche Diagnostics) and frozen in liquid nitrogen before resuspension in extraction buffer and homogenization with zirconia beads (four cycles, 6,000 × 2; 20 s; Precellys24 Homogenizer with Cryolys Cooling Unit). Supernatant was collected after 15 min of centrifugation at 4°C. The total protein concentration was measured using the Pierce 660 nm Protein assay (Thermo Fisher Scientific).

Phosphoproteome

Protein pellets were re-suspended in 6 M urea and 2 M thio-urea in 10 mM HEPES buffer with a Bioruptor, digested in solution with Lys-C and trypsin enzymes, and then enriched with phosphopeptide with the use of TiO₂ beads before LC-MS/MS.

Statistical Analysis

MS raw files were analyzed with MaxQuant version 1.5.2.8 software and the UniProt reference proteome database for C. elegans. Protein and peptide identifications were controlled by a decoy-database approach FDR estimation to 0.01. GO annotations were imported on the basis of UniProt entries with Perseus. Statistical analyses (correlation calculations and visualization) were performed in the software package Rstudio version 0.99.489 (2009–2015) and network analysis in Cytoscape with WormBase as the reference network. Significantly altered protein quantities were determined by two-sided t tests and correction for multiple testing via estimation of the FDR to 5% with a permutation-based algorithm (number of permutations = 500, fudge factor s0 = 0.1). Significant differences between developmental stages were assessed by two-tailed t tests. For lipid classes, independent two-group t tests were applied.

Lipid Analysis

Lipids were extracted from 500,000 L1 larvae. Triacyglycerols and free FAs were quantified by analytical TLC. Relative amounts of SLs were determined by liquid chromatography coupled to electrospray ionization tandem MS (LC-ESI-MS/MS). Glycerophospholipids were analyzed by ESI-MS/MS with direct infusion of the lipid extract (Shotgun Lipidomics).

Protein Expression

Worm pellet was sonicated 2 × 5 s at 40% power and boiled 5 min at 94°C in Laemmli buffer, separated on SDS-PAGE gels (4%–12% resolving gel; Invitrogen), transferred to a nitrocellulose membrane (Protran, 0.2 μm; Whatman) and incubated with primary antibodies against GFP (JL-8 Living Colors; Clontech) and tubulin (mouse monoclonal; Sigma-Aldrich) in a 10% Roti-Block (Roth) and imaged with the Odyssey Infrared Imaging System (Li-Cor Bioscience).

Author Contributions

D.E. and B.S. conceived the study. D.E, H.N., V.B., L.C.-R., M.M.M., and S.B. performed the experiments. D.E., H.N., S.B., and M.K. analyzed the data. D.E., H.N., and B.S. wrote the manuscript.

Published: August 29, 2017

Accession Numbers

The extended protein-protein interaction network is available at http://bifacility.uni-koeln.de/schumacher/suppdata. The accession number for the MS proteomics data reported in this paper is ProteomeXchange Consortium: PXD005649.

Supplemental Information

Table S1. Proteins Quantified in xpc-1;csb-1 Double Mutants at FDR < 1%, Related to Figure 1

More than 5000 proteins were quantified between UV and untreated conditions at least in two out of three biological replicates.

Table S2. One-Dimensional Analysis of Human and C. elegans GO Categories Found Differentially Regulated upon UV Treatment, Related to Figure 1D

The main GO categories found regulated are highlighted in different colours in a range from 1 to 6.

Table S3. The Two Tables Recapitulate the Most Represented Clusters of Proteins that Were Found Up- and/or Downregulated at the Proteome (xpc-1;csb-1 Double Mutants) and at the Transcriptome (xpa-1 Mutants) Level after UV Treatment, Related to Figure 3A

Red arrows refer to upregulation while blue arrows refer to downregulation.

Table S4. Phosphoroteome Data Set from xpc-1;csb-1 Double Mutants upon Each Treatment (Untreated, UV Treated, and Starvation), Related to Figure 5

Among the 7000 detected phosphosites, we identified more than 3000 significantly modulated in response to UV treatment.